app:
  description: 漫画生成工具验证 - 从小说文本到分镜图片提示词
  icon: 🎨
  icon_background: '#E3F2FD'
  mode: workflow
  name: 漫画生成验证工作流
  use_icon_as_answer_icon: false
kind: app
version: 0.1.2
workflow:
  conversation_variables: []
  environment_variables: []
  features:
    file_upload:
      enabled: false
    opening_statement: '请输入小说段落，我会帮你生成漫画分镜提示词。'
    retriever_resource:
      enabled: false
    sensitive_word_avoidance:
      enabled: false
    speech_to_text:
      enabled: false
    suggested_questions_after_answer:
      enabled: false
    text_to_speech:
      enabled: false
  graph:
    edges:
    - data:
        isInIteration: false
        sourceType: start
        targetType: llm
      id: 1740748800000-source-1740748800001-target
      source: '1740748800000'
      sourceHandle: source
      target: '1740748800001'
      targetHandle: target
      type: custom
      zIndex: 0
    - data:
        isInIteration: false
        sourceType: llm
        targetType: llm
      id: 1740748800001-source-1740748800002-target
      source: '1740748800001'
      sourceHandle: source
      target: '1740748800002'
      targetHandle: target
      type: custom
      zIndex: 0
    - data:
        isInIteration: false
        sourceType: llm
        targetType: answer
      id: 1740748800002-source-1740748800003-target
      source: '1740748800002'
      sourceHandle: source
      target: '1740748800003'
      targetHandle: target
      type: custom
      zIndex: 0
    nodes:
    - data:
        desc: ''
        selected: false
        title: 开始
        type: start
        variables:
        - label: 小说原文
          max_length: 2000
          options: []
          required: true
          type: paragraph
          variable: input_novel
      height: 89
      id: '1740748800000'
      position:
        x: 92
        y: 292
      selected: false
      sourcePosition: right
      targetPosition: left
      type: custom
      width: 244
    - data:
        context:
          enabled: false
        desc: 润色和增强小说文本描述
        model:
          completion_params:
            max_tokens: 2000
            temperature: 0.7
          mode: chat
          name: deepseek-r1
          provider: deepseek
        prompt_template:
        - id: system-001
          role: system
          text: |
            你是一个专业的小说编辑。你的任务是将用户提供的小说段落进行润色和增强，使其更适合转换为视觉分镜。

            要求：
            1. 保持原有情节和人物设定不变
            2. 增强场景描述、角色动作和情感表达
            3. 让文本更具画面感和戏剧张力
            4. 输出保持中文，长度控制在原文本的1.2-1.5倍
        - id: user-001
          role: user
          text: '{{#1740748800000.input_novel#}}'
        selected: false
        title: 小说编辑AI
        type: llm
        variables: []
      height: 98
      id: '1740748800001'
      position:
        x: 384
        y: 292
      selected: false
      sourcePosition: right
      targetPosition: left
      type: custom
      width: 244
    - data:
        context:
          enabled: false
        desc: 将叙述文本转换为分镜提示词
        model:
          completion_params:
            max_tokens: 3000
            temperature: 0.6
          mode: chat
          name: deepseek-r1
          provider: deepseek
        prompt_template:
        - id: system-001
          role: system
          text: |
            你是一个专业的漫画分镜师。请将叙述文本转换为3-5个分镜描述，每个分镜包含视觉元素和图片生成提示词。

            输出格式要求（Markdown）：
            # 分镜分析

            ## 场景概览
            [简要描述整体场景]

            ## 分镜详情

            ### 分镜1：[标题]
            **场景描述：** [详细视觉描述]
            **角色动作：** [主要角色动作]
            **情感氛围：** [情感色彩]
            **图片提示词：** [用于图片生成的提示词，包含风格、构图等]

            ### 分镜2：[标题]
            ...

            ## 风格建议
            [推荐的图片生成风格，如：动漫风格、写实插画、水彩等]

            注意：图片提示词要具体、详细，包含构图、灯光、风格等要素。
        - id: user-001
          role: user
          text: '{{#1740748800001.text#}}'
        selected: false
        title: 分镜AI
        type: llm
        variables: []
      height: 98
      id: '1740748800002'
      position:
        x: 676
        y: 292
      selected: false
      sourcePosition: right
      targetPosition: left
      type: custom
      width: 244
    - data:
        answer: '{{#1740748800002.text#}}'
        desc: 输出分镜提示词
        selected: false
        title: 分镜提示词输出
        type: answer
        variables: []
      height: 102
      id: '1740748800003'
      position:
        x: 968
        y: 292
      selected: false
      sourcePosition: right
      targetPosition: left
      type: custom
      width: 244